Conventionally, blue/violet LED's have been made using InGaN alloys to produce the layers thereof. Problems encountered when fabricating optical sources using high dislocation density III-Nitride materials, such as InGaN alloys, was first addressed during the development of LED's using the InxGa1-xN alloy system. It was found that the material properties of InGaN alloys change as the amount of indium in the alloy is increased. With the proper growth conditions, however, it was discovered that material could be grown in which the indium did not incorporate uniformly throughout the InGaN layer (i.e., the material has areas of high and low concentrations of indium spread throughout). These compositional fluctuations, also known as localized inhomogeneities, result in carrier localization and lead to an enhancement in the radiative efficiency despite the high dislocation density.
The discovery of the effects of the localized inhomogeneities enabled the development of commercially successful blue InGaN-based LEDs and laser diodes (LDs). See P. Waltereit, H. Sato, C. Poblenz, D. S. Green, J. S. Brown, M. Mclaurin, T. Katona, S. P. DenBaars, J. S. Speck, J. H. Liang, M. Kato, H. Tamura, S. Omori, and C. Funaoka, Appl. Phys. Lett. 84, 2748 (2004); S. Nakamura, M. Senoh, N. Iwasa, and S. Nagahama, Appl. Phys. Lett. 67, 1868 (1995); S. Nakamura, J. Vac. Sci. Technol. A 13, 705 (1995); S. Nakamura, T. Mukai, M. Senoh, J. Appl. Phys. 76, 8189, (1994), each of which are incorporated herein by reference in their entireties.
Specifically, it has been reported that the intense red-shifted photoluminescence (PL) peaks observed in InGaN alloys at room temperature result from the recombination of excitons localized at potential minima originating from large compositional fluctuations. See S. Chichibu, T. Azuhata, T. Sota, and S. Nakamura, Appl. Phys. Lett. 70, 2822 (1997); S. Chichibu, K. Wada, and S. Nakamura, Appl. Phys. Lett. 71, 2346 (1997); Y. Sun, Y. Cho, E. K. Suh, H. J. Lee, R. J. Choi, and Y. B. Hahn, Appl. Phys. Lett. 84, 49 (2004), each of which are incorporated herein by reference in their entireties. Unfortunately, InGaN alloys are not suitable for use in UV LED's at wavelengths shorter than 365 nanometers (nm).
A similar localization effect resulting from the creation of localized inhomogeneities has been reported for quaternary InAlGaN alloys having an InN mole fraction of up to 20% and an AlN mole fraction of up to 60%, corresponding to light emission at wavelengths shorter than 365 nm. See E. Monroy, N. Gogneau, F. Enjalbert, F. Fossard, D. Jalabert, E. Bellet-Amalric, Le Si Dang, and B. Daudin, J. Appl. Phys. 94, 3121 (2003); Mee-Yi Ryu, C. Q. Chen, E. Kuokstis, J. W. Yang, G. Simin, and M. Asif Khan, Appl. Phys. Lett. 80, 3730 (2002); H. Hirayama, A. Kinoshita, T. Yamabi, Y. Enomoto, A. Hirata, T. Araki, Y. Nanishi, and Y. Aoyagi, Appl. Phys. Lett. 80, 207 (2002); C. H. Chen, Y. F. Chen, Z. H. Lan, L. C. Chen, K. H. Chen, H. X. Jiang, and J. Y. Lin, Appl. Phys. Lett. 84, 1480 (2004), each of which are incorporated herein by reference in their entireties. The carrier localization observed in such InAlGaN quaternary materials is also attributed to indium segregation therein, resulting in enhanced luminescence at a peak wavelength significantly red-shifted (150 to 300 meV) with respect to the band edge.
The use of aluminum gallium nitride (AlxGa1-xN), as opposed to InAlGaN, is preferred as the base material for manufacturing UV LED devices due to the difficulty in growing device quality quaternary material (InAlGaN). AlGaN has, therefore, become the most important material for use in the development of ultraviolet semiconductor optical sources operating at wavelengths between 260 to 360 nm due to its tunable bandgap from 3.4 eV to 6.2 eV. Such sources have many useful military and commercial applications, including water purification, phosphor based white light sources, high-density data storage, bioagent detection and non-line of sight (NLOS) covert communication. See G. A. Shaw, A. M. Siegel, J. Model, and N. Nischan, SPIE Defense & Security Symposium, Vol. 5417, (2004), and E. Radkov, R. Bompiedi, A. M. Srivastava, A. A. Setlur, and C. A. Becket, Proc. SPIE Int. Soc. Opt. Eng., Vol. 5187, 176, (2004), both of which are incorporated herein by reference in their entireties.
A major problem in manufacturing practical III-Nitride based sources, such as AlGaN alloys, is the lack of a native substrate for the homoepitaxial growth of epilayers. Consequently, most devices are deposited heteroepitaxially on lattice mismatched substrates, like sapphire or SiC, with the resulting layer quality limited by the high density of threading dislocations produced therein. Such threading dislocations present in AlGaN alloys deposited heteroepitaxially on lattice mismatched substrates are associated with non-radiative recombination centers that compete with radiative recombination paths, thus reducing the radiative recombination efficiency of these materials. One method for reducing the number of dislocations in III-nitride materials deposited by molecular beam epitaxy is to employ III-Nitride templates as a substrate that consists of an (In)(Ga)(Al)N thin film deposited upon a substrate, commonly sapphire or SiC, by a high temperature growth process (>1000° C.) such as metalorganic chemical vapor deposition or hydride vapor phase epitaxy.
With regard to optical sources, these threading dislocations present in AlGaN alloys deposited heteroepitaxially on lattice mismatched substrates greatly reduce the wall plug efficiency (i.e., the ratio of the power of the light emitted versus the electrical power applied). Currently, the best reported UV light emitting diodes (LED) have very low wall plug efficiencies, ˜1% or below, with the remaining input power being converted to heat. See A. Khan, Light-Emitting Diodes: Research, Manufacturing, and Applications VII, SPIE Vol. 4996, (2003); T. M. Katona, T. Margalith, C. Moe, M. C. Schmidt, S. Nakamura, J. S. Speck, S. P. DenBaars, Third International Conference on Solid State Lighting, SPIE Vol. 5187, (2004); K. B. Nam, J. Li, M. L. Nakarmi, J. Y. Lin, and H. X. Jiang, Appl. Phys. Lett. 84, 5264 (2004); S. Wieczorek, W. W. Chow, S. R. Lee, A. J. Fisher, A. A. Allerman, and M. H. Crawford, Appl. Phys. Lett. 84, 4899 (2004); M. H. Crawford, A. A. Allerman, A. J. Fisher, K. H. A. Bogart, S. R. Lee, R. J. Kaplar, W. W. Chow, and D. M. Follstaedt, Light-Emitting Diodes: Research, Manufacturing, and Applications VIII, SPIE Vol. 5366, (2004); A. Chitnis, J. Sun, V. Mandavilli, R. Pachipulusu, S. Wu, M. Gaevski, V. Adivarahan, J. P. Zhang, M. Asif Khan, A. Sarua, and M. Kuball, Appl. Phys. Lett. 81, 3491 (2002); V. Adivarahan, S. Wu, J. P. Zhang, A. Chitnis, M. Shatalov, V. Mandavilli, R. Gaska, and M. Asif Khan, Appl. Phys. Lett. 84, 4762 (2004); V. Adivarahan, W. H. Sun, A. Chitnis, M. Shatalov, S. Wu, H. P. Maruska, and M. Asif Khan, Appl. Phys. Lett. 85, 2175 (2004), each of which are incorporated herein by reference in their entireties.
The low wall plug efficiency further results in excessive heating of the UV LED material, which reduces the lifetime thereof, and may change the spectral and power output properties of UV LEDs. For these reasons, such UV LED devices are commonly pulsed in operation, such that the duty cycle is kept low enough to reduce undesirable heating effects on the UV LED material. Extracting the heat from these devices can also be challenging since the most commonly used substrate, sapphire, is a poor thermal conductor at room temperature. This difficulty in heat extraction requires elaborate packaging of the UV LED material, such as including flip-chip bonding of the final device to a thermally conductive substrate to help conductively dissipate the heat. These devices then emit light through the sapphire substrate, and are referred to as “back-emitters”.
The red-shifts in the enhanced luminescence seen in In(Al)GaN quaternary alloys are much larger than those typically reported for AlxGa1-xN alloys of comparable Al content (˜10 to 50 meV). These AlxGa1-xN alloys are referred to herein as band edge AlGaN alloys. Accordingly, it has been found that the large compositional fluctuations, i.e., localized inhomogeneities resulting in the desirable carrier localization that leads to an enhancement in the radiative efficiency, obtained using the current growth methods of InxGa1-xN alloys do not occur when using such methods to grow AlxGa1-xN alloys needed for the production of UV LEDs. In fact, it has been previously believed that AlGaN material systems cannot have unstable mixing regions (i.e., localized inhomogeneities) that would result in phase separation, as is the case for InGaN materials. See T. Matsuoka, Calculation of unstable mixing region in wurtzite InGaAlN, Applied Physics Letters 71, 105 (1997).
These smaller red-shifts found in the band edge AlxGa1-xN alloys of comparable Al content (˜10 to 50 meV) are often attributed to emission from bandtails associated with small alloy fluctuations and structural disorder. See Y. H. Cho, G. H. Gainer, J. B. Lam, W. Yang, W. Jhe, and J. J. Song, Phys. Rev. B 61, 7203 (2000); A. Bell, S. Srinivasan, C. Plumlee, H. Omiya, F. A. Ponce, J. Christen, S. Tanaka, A. Fujioka, and Y. Nakagawa, J. Appl. Phys. 95, 4670 (2004); H. S. Kim, R. A. Mair, J. Li, J. Y. Lin, and H. X. Jiang, Appl. Phys. Lett. 76, 1252 (2000), each of which are incorporated herein by reference in their entireties. While AlxGa1-xN on sapphire possesses long-lived low temperature photoluminescence (PL) on the order of 250-600 picoseconds (ps), a significant reduction in PL lifetime, to less than 20 ps at room temperature, is generally observed due to thermally activated trapping and non-radiative recombination at defect sites in the AlxGa1-xN alloy.
These short room temperature lifetimes found in UV LED's comprising band edge AlxGa1-xN on sapphire correspond to low wall plug efficiencies, as discussed above. Without the availability of lattice matched substrates with reduced defect density, or the advantages of compositional fluctuations (i.e., localized inhomogeneities) found in InGaN alloys made using the conventional methods of production, production of efficient UV LED using band edge AlGaN alloys has been found to be practically impossible.
Most conventional UV LED's are based on the same basic device structure. The substrate preparation and the nucleation steps for epilayer growth, however, are dependent upon the chemistry of the substrate used. These processes, as well as the subsequent buffer layers deposited, greatly affect the number of defects in the material. For example, in a conventional UV LED device structure, a thick silicon-doped n-type AlxGa1-xN layer is grown to function as the bottom current-spreading layer of the UV LED device. For a back-emitter device, as described above, the AlN mole fraction of this current-spreading layer is made large enough to ensure that the current-spreading layer is transparent to the light emitted from the active region layer.
Next, a multiple quantum well (MQW) active region layer is deposited upon the current-spreading layer, with the number and thickness of the multiple quantum wells formed within the MQW layer depending on the particular device. This MQW layer is then generally capped with an AlGaN electron blocking layer, the AlGaN electron blocking layer optionally being doped p-type. Next, a p-type GaN contact layer is formed upon the AlGaN electron blocking layer. The p-type GaN contact layer is required due to the difficulties in doping AlGaN p-type.
The p-type GaN contact layer may optionally be replaced by many different AlGaN p-type layers depending on the particular device. If the MQW active region layer is replaced, for example, by an AlGaN layer with one aluminum percentage, then the device is called a double-heterostructure (DH) LED. While this conventional UV LED device structure is simple to grow, it is not possible to produce efficient DH-UV LEDs using the above-described method, due to the susceptibility of the active region layer to the non-radiative recombination centers typically found in these materials, polarization fields that tend to separate the electron hole pairs, and the lack of quantum confinement to improve electron-hole wave function overlap. The latter two reasons lead to a longer radiative lifetime in the bulk active layers that does not compete favorably with the non-radiative processes.
One method for depositing III-Nitride films is a technique called molecular beam epitaxy (MBE). This technique is a thin film deposition process in which thermal beams of atoms or molecules react on the clean surface of a substrate, held at high temperatures and under ultra-high vacuum conditions. The material sources for III-Nitride semiconductors traditionally consist of solid elemental constituent sources of gallium (Ga), aluminum (Al) and indium (In), that evaporate from the melt, as well as n- and p-type dopant sources silicon (Si) and magnesium (Mg) that sublimate. Since molecular nitrogen (N2) does not crack on the substrate surface, plasma sources are employed to generate active nitrogen species from this gas source. Accordingly, this molecular beam epitaxy technique is commonly referred to as plasma-assisted molecular beam epitaxy (PA-MBE).
In this method, crystal growth is performed in a chamber with ultra-high vacuum (<10−9 Torr) base pressure (not including the partial pressure of source gases) that allows for the deposition of high-purity materials at lower substrate temperatures than typically employed by vapor deposition techniques. This environment allows for the use of a number of in-situ probes, one of the most common being reflection high energy electron diffraction (RHEED). In this technique, a high energy (10 to 30 keV) electron beam is directed at the substrate at a grazing angle (˜1°) with respect to the surface.
Since the penetration depth of the beam is only a few atomic layers, the crystalline surface acts as a two-dimensional grating that diffracts electrons. The diffraction pattern can be observed on a fluorescent screen located inside the MBE chamber. A film having a smooth surface, as illustrated in the SEM photograph of the conventional GaN layer grown at a high Ga flux shown in FIG. 1(a), is expected to have a pattern consisting of a series of streaks perpendicular to the substrate surface, as illustrated in FIG. 1(b), while a rough or polycrystalline film, as illustrated in the SEM photograph of a GaN layer grown at a low Ga flux, as illustrated in FIG. 2(a), would have a spotty pattern, as illustrated in FIG. 2(b).
Extensive work has been performed in this field on the growth of GaN using PA-MBE. It is known that two of the most important growth parameters for the deposition of high quality GaN epilayers are the substrate temperature and the group-III/group-V atomic flux ratio. Unlike the deposition of other III-V compounds by MBE, it is well known that high quality, smooth GaN films are deposited under a Ga-rich (N-limited) growth regime. This is due to the highly reactive nature of the active nitrogen species generated by the plasma source that results in very short adatom diffusion lengths.
Suitable parameter space for depositing GaN has been reported upon, and three distinct growth regimes identified, referred to as the N-stable, intermediate Ga stable and Ga droplets, as illustrated in FIG. 3 (See B. Heying, R. Averbeck, L. F. Chen, E. Haus, H. Reichert, J. S. Speck, “Control of surface morphologies using plasma-assisted molecular beam epitaxy”, Journal of Applied Physics, 88, 1855 (2000)). The boundary between the N-stable and the intermediate Ga stable regimes constitutes conditions that are nearly stoichiometric where the active nitrogen atom flux equals the Ga atom flux, and was determined by examining the dependence of the growth rate of the GaN film on the Ga atom flux.
The growth rate of GaN has been found to be independent of Ga atom flux (or Ga atom arrival rate) only within the intermediate Ga stable and Ga-rich regimes. Heavily Ga-rich growth conditions results in the accumulation of liquid Ga metal droplets on the surface of the film. Increasing the substrate temperature during growth has been observed to widen the range of Ga atom fluxes that result in intermediate Ga stable for a given active nitrogen flux.
As illustrated in FIG. 3, and as mentioned above, GaN film can be grown under three distinct growth regimes, i.e., the N-stable, intermediate Ga stable and Ga droplets. The surface morphology and optical properties of the GaN film have all been observed to depend upon the regime (method) of growth employed. Heying, et al. have reported on the surface morphologies of GaN films deposited within each regime, and have observed that films deposited under N— stable conditions had morphologies that were rough and heavily pitted, while films grown within the intermediate regime had large flat areas between large irregularly shaped pits. In contrast, films deposited in the Ga droplet regime had atomically flat surfaces.
The present inventors have investigated the impact of III/V flux ratio on the continuous wave photoluminescence observed from GaN films, and have found that films grown near stoichiometric conditions (i.e., where III/V flux ratio ˜1) had significantly more intense band-edge luminescence than films grown at higher III/V flux ratio. Similarly, the present inventors observed that films deposited near stoichiometric conditions had significantly longer photogenerated carrier lifetimes. These results are believed to be attributable to a high density of deep trap states in films deposited with higher III/V flux ratio.
In contrast to GaN, comparatively less has been reported on the growth of AlxGa1-xN films by PA-MBE, which are useful in producing UV LED's. The present inventors have discovered that the AlN mole fraction of this ternary alloy (AlxGa1-xN) is dependent upon the ratio of the Al/Ga atomic flux, the ratio of the total group III/group V atomic flux, as well as the substrate temperature. It has been previously observed that the AlN mole fraction of AlGaN alloys, deposited under constant Al/Ga atomic flux ratio, increases with increasing III/V flux ratio. This has been attributed to the preference for Al to incorporate into the film over Ga, due to the greater thermal stability of AlN over GaN. Presently, however, a growth diagram (methodology for production) for AlGaN alloys, comparable to the one developed for GaN as discussed above, has not been reported.
In view of the above, it is an object of the present invention to provide an AlGaN composition for use as an active region layer of a UV light emitting device, in which the active region layer's composition has localized inhomogeneities therein, resulting in carrier localization and enhanced radiative efficiency, and a UV light emitting device containing same.
It is a further object of the present invention to provide a method of manufacturing an AlGaN composition for use as an active region layer of a UV light emitting device, in which the active region layer has localized inhomogeneities formed therein, resulting in carrier localization and enhanced radiative efficiency, and a UV light emitting device containing same, as described above.